INDEX

Explanations

phrases related to completion or achievement of tasks

oai_token-act-pair · gpt-3.5-turbo

phrases indicating potential actions or occurrences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

=~

-0.62

azel

-0.61

 alike

-0.59

anything

-0.59

 likewise

-0.59

both

-0.58

also

-0.57

etc

-0.57

auga

-0.57

robat

-0.56

POSITIVE LOGITS

 marginally

0.95

 spor

0.93

 insofar

0.90

ONE

0.85

 fraction

0.76

 peripher

0.76

 temporarily

0.74

 curs

0.73

 fleeting

0.70

 briefly

0.69

Activations Density 0.242%

phrases related to completion or achievement of tasks

phrases indicating potential actions or occurrences

No Comments

No Known Activations