INDEX

Explanations

references to feeling proud or expressing pride

oai_token-act-pair · gpt-3.5-turbo

expressions of pride

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

PF

-0.79

apter

-0.73

itamin

-0.68

 bite

-0.62

CPU

-0.62

vati

-0.62

intestinal

-0.61

 Webster

-0.61

 mosquit

-0.60

 halting

-0.60

POSITIVE LOGITS

edIn

0.91

hon

0.84

edly

0.81

roud

0.77

ledged

0.75

kson

0.73

sburg

0.72

faced

0.71

 proud

0.71

 citiz

0.70

Activations Density 0.013%

references to feeling proud or expressing pride

expressions of pride

No Comments

No Known Activations