INDEX

Explanations

phrases related to participating in events or activities

oai_token-act-pair · gpt-3.5-turbo

references to political parties or affiliations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ħĭ

-0.94

¥µ

-0.80

éĹĺ

-0.77

OME

-0.77

hirt

-0.76

hower

-0.76

æ©Ł

-0.74

doms

-0.74

å£«

-0.73

;;;;;;;;;;;;

-0.71

POSITIVE LOGITS

rot

0.96

amount

0.89

agraph

0.88

allel

0.86

rots

0.83

anoia

0.79

aged

0.79

 ILCS

0.79

ret

0.78

icularly

0.78

Activations Density 0.009%

phrases related to participating in events or activities

references to political parties or affiliations

No Comments

No Known Activations