INDEX

Explanations

terms related to different sections or components within a larger structure or system

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "each" or terms that emphasize distribution or individuality across multiple items

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

aukee

-0.65

DERR

-0.62

 Newsletter

-0.62

 Things

-0.61

mares

-0.61

tymology

-0.57

wives

-0.56

dylib

-0.56

Ire

-0.56

Demand

-0.56

POSITIVE LOGITS

 individually

1.16

 separately

0.92

ieth

0.75

adelphia

0.75

isan

0.71

atic

0.70

 respective

0.67

 contributes

0.67

tick

0.66

 vying

0.66

Activations Density 0.091%

terms related to different sections or components within a larger structure or system

occurrences of the word "each" or terms that emphasize distribution or individuality across multiple items

No Comments

No Known Activations