INDEX

Explanations

phrases related to cutting or trimming actions

oai_token-act-pair · gpt-3.5-turbo

phrases related to divisions or reductions in context

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

SPONSORED

-1.03

ixtures

-0.75

SIGN

-0.70

ITNESS

-0.69

wick

-0.69

eton

-0.68

Own

-0.67

âĢ¢âĢ¢âĢ¢âĢ¢

-0.67

CLASS

-0.67

MAP

-0.66

POSITIVE LOGITS

 corners

0.74

erous

0.70

umb

0.69

 slack

0.67

 manpower

0.67

 brackets

0.65

 overhead

0.63

thro

0.62

 wast

0.62

 redundant

0.61

Activations Density 0.138%

phrases related to cutting or trimming actions

phrases related to divisions or reductions in context

No Comments

No Known Activations