INDEX

Explanations

specific values or quantities mentioned in a technical context

oai_token-act-pair · gpt-3.5-turbo

phrases indicating possession or ownership

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

 prosecuting

-0.76

ensing

-0.70

eem

-0.69

bart

-0.64

—-

-0.61

 territ

-0.61

zone

-0.59

icking

-0.59

 wounding

-0.59

uating

-0.58

POSITIVE LOGITS

 been

1.27

 undergone

1.07

kell

0.99

been

0.98

 Been

0.94

drawn

0.91

stood

0.87

hed

0.85

 gotten

0.85

 gone

0.81

Activations Density 0.435%

specific values or quantities mentioned in a technical context

phrases indicating possession or ownership

No Comments

No Known Activations