INDEX

Explanations

mentions of being physically covered or hidden by something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "covered."

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

mage

-0.69

rw

-0.64

Sa

-0.64

assuming

-0.62

cgi

-0.61

 Sabha

-0.61

rained

-0.59

lua

-0.59

 correctness

-0.59

enc

-0.58

POSITIVE LOGITS

 Coverage

0.79

krit

0.79

iday

0.78

Cover

0.77

ummer

0.72

 COVER

0.71

bys

0.71

utical

0.71

ãĤ¹

0.70

alls

0.70

Activations Density 0.018%

mentions of being physically covered or hidden by something

instances of the word "covered."

No Comments

No Known Activations