INDEX

Explanations

words related to deception or trickery

oai_token-act-pair · gpt-3.5-turbo

instances of the word "deceive" and its variations or related forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Kenobi

-0.75

 Skydragon

-0.73

isations

-0.69

ounces

-0.68

izational

-0.67

mort

-0.66

asking

-0.66

izations

-0.66

unci

-0.66

ested

-0.65

POSITIVE LOGITS

utical

1.00

les

0.90

pter

0.90

lled

0.85

lla

0.84

ce

0.83

lectic

0.80

rette

0.79

lling

0.79

Activations Density 0.037%

words related to deception or trickery

instances of the word "deceive" and its variations or related forms

No Comments

No Known Activations