INDEX

Explanations

words related to trickery or deception

oai_token-act-pair · gpt-3.5-turbo

mentions of the term "fool" and its variations

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

 Parenthood

-0.81

lined

-0.63

accompan

-0.63

ials

-0.61

orney

-0.60

capacity

-0.60

Balt

-0.59

 disadvantage

-0.58

battle

-0.58

lining

-0.58

POSITIVE LOGITS

ery

0.97

hard

0.93

pas

0.90

sonian

0.84

ulent

0.83

eries

0.82

ument

0.81

usional

0.79

ibility

0.78

hemer

0.78

Activations Density 0.055%

words related to trickery or deception

mentions of the term "fool" and its variations

No Comments

No Known Activations