INDEX

Explanations

the word "know"

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "you know" or related expressions indicating conversational engagement

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

utenberg

-0.81

anmar

-0.79

aez

-0.67

ersive

-0.65

jug

-0.64

arthed

-0.64

acco

-0.63

rang

-0.63

luaj

-0.62

otom

-0.62

POSITIVE LOGITS

terday

0.82

 WHAT

0.81

 what

0.79

how

0.75

why

0.72

 whats

0.70

what

0.65

ledged

0.65

ees

0.61

 imagine

0.61

Activations Density 0.037%

the word "know"

instances of the phrase "you know" or related expressions indicating conversational engagement

No Comments

No Known Activations