INDEX

Explanations

statements or phrases indicating falsehood or deception

oai_token-act-pair · gpt-3.5-turbo

statements related to falsehoods, lies, or inaccuracies

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 daunting

-0.72

 restless

-0.68

cellaneous

-0.67

 teased

-0.65

 melancholy

-0.64

 goodbye

-0.63

 finally

-0.63

 someday

-0.62

aunted

-0.61

 farewell

-0.61

POSITIVE LOGITS

facts

0.89

ItemImage

0.87

 factual

0.83

Claim

0.83

Fact

0.83

 merely

0.80

 Actually

0.80

 neither

0.79

 untrue

0.78

 misrepresent

0.77

Activations Density 0.888%

statements or phrases indicating falsehood or deception

statements related to falsehoods, lies, or inaccuracies

No Comments

No Known Activations