INDEX

Explanations

references to legal documents or regulatory agencies

oai_token-act-pair · gpt-3.5-turbo

occurrences of brackets or bracket-like characters

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

hog

-0.81

 lemon

-0.71

 consumption

-0.69

 uncertain

-0.65

 maximum

-0.65

 seams

-0.65

 stag

-0.63

 beet

-0.63

Orn

-0.62

 raspberry

-0.62

POSITIVE LOGITS

Pg

1.34

âĢ¦]

1.26

...]

1.23

sic

1.08

paragraph

1.04

Footnote

0.99

interstitial

0.97

src

0.96

?]

0.92

](

0.88

Activations Density 0.019%

references to legal documents or regulatory agencies

occurrences of brackets or bracket-like characters

No Comments

No Known Activations