INDEX

Explanations

phrases related to physical force or motion

oai_token-act-pair · gpt-3.5-turbo

terms related to disciplinary actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

igmatic

-0.78

gow

-0.78

oret

-0.73

inav

-0.70

zynski

-0.68

NTS

-0.66

Import

-0.66

mund

-0.64

 Diff

-0.64

tex

-0.64

POSITIVE LOGITS

 leash

0.89

 blaster

0.79

 nails

0.78

 darts

0.78

 away

0.77

knife

0.76

gered

0.73

 levers

0.73

 thumbs

0.73

 vines

0.72

Activations Density 0.063%

phrases related to physical force or motion

terms related to disciplinary actions

No Comments

No Known Activations