INDEX

Explanations

verbs of action or consequence, specifically related to personal experiences

oai_token-act-pair · gpt-3.5-turbo

expressions of permission or agency

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 meanwhile

-0.63

 however

-0.62

Rai

-0.61

 though

-0.61

 straw

-0.59

LD

-0.59

 heats

-0.56

 Hearth

-0.56

 Crunch

-0.56

Hai

-0.55

POSITIVE LOGITS

aff

0.67

ãĥīãĥ©

0.66

etheless

0.64

interstitial

0.63

*.

0.62

 oneself

0.60

'.

0.59

BILITIES

0.59

ãĥĻ

0.58

uana

0.58

Activations Density 0.870%

verbs of action or consequence, specifically related to personal experiences

expressions of permission or agency

No Comments

No Known Activations