INDEX

Explanations

words or phrases related to uncovering or discovering information

oai_token-act-pair · gpt-3.5-turbo

references to clues or hints related to mysteries or investigations

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

lav

-0.74

rifice

-0.74

rior

-0.71

kus

-0.69

sburgh

-0.68

ategory

-0.65

 Turks

-0.64

rik

-0.64

rie

-0.64

roc

-0.63

POSITIVE LOGITS

 clue

1.12

 hint

0.95

 clues

0.90

 glean

0.78

 hints

0.72

hole

0.71

hig

0.69

wcs

0.67

 detector

0.67

hooting

0.66

Activations Density 0.022%

words or phrases related to uncovering or discovering information

references to clues or hints related to mysteries or investigations

No Comments

No Known Activations