INDEX

Explanations

references to badges, which can symbolize accomplishment, identity, or authority

oai_token-act-pair · gpt-3.5-turbo

references to badges and symbols of honor or status

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

noon

-0.94

moving

-0.80

theless

-0.78

lihood

-0.72

TION

-0.67

 Islands

-0.66

ted

-0.66

frey

-0.66

scill

-0.65

quished

-0.65

POSITIVE LOGITS

 badges

1.19

 badge

1.17

 insign

1.01

 Badge

0.84

 holder

0.79

 bearer

0.78

 fide

0.77

 commemor

0.76

holder

0.75

holders

0.75

Activations Density 0.020%

references to badges, which can symbolize accomplishment, identity, or authority

references to badges and symbols of honor or status

No Comments

No Known Activations