INDEX

Explanations

words related to important structures or supporting elements

oai_token-act-pair · gpt-3.5-turbo

terms related to structural support or foundational concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

aps

-0.91

Attempts

-0.79

vertisements

-0.79

avers

-0.77

 therape

-0.76

aped

-0.76

imar

-0.76

umatic

-0.75

ivan

-0.74

aver

-0.73

POSITIVE LOGITS

 backbone

1.50

layer

0.96

SourceFile

0.87

 guts

0.78

xual

0.74

REDACTED

0.73

 fibre

0.72

Connector

0.72

beard

0.72

 bones

0.71

Activations Density 0.010%

words related to important structures or supporting elements

terms related to structural support or foundational concepts

No Comments

No Known Activations