INDEX

Explanations

words related to instructions or procedures

oai_token-act-pair · gpt-3.5-turbo

words and phrases related to specific outcomes or results

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

pta

-0.52

edo

-0.51

edom

-0.50

iries

-0.49

 confir

-0.47

TI

-0.47

ãĤ§

-0.46

itiveness

-0.46

ĭ

-0.45

FACE

-0.45

POSITIVE LOGITS

in

1.39

 therein

1.00

In

0.94

In

0.92

in

0.90

IN

0.89

inant

0.71

ÃŃn

0.70

IN

0.66

inic

0.58

Activations Density 1.707%

words related to instructions or procedures

words and phrases related to specific outcomes or results

No Comments

No Known Activations