INDEX

Explanations

contradictions between different pieces of information or statements

oai_token-act-pair · gpt-3.5-turbo

statements and phrases that express contradiction or refutation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

emetery

-0.79

NetMessage

-0.76

anny

-0.75

ussed

-0.71

 actionGroup

-0.67

itizens

-0.66

Eye

-0.64

ammy

-0.64

umbered

-0.62

hene

-0.62

POSITIVE LOGITS

 corrobor

0.98

 contradict

0.97

 substant

0.93

 dispro

0.89

 debunk

0.84

 assertions

0.84

 refute

0.83

 hypotheses

0.78

 debunked

0.76

 unfounded

0.76

Activations Density 0.047%

contradictions between different pieces of information or statements

statements and phrases that express contradiction or refutation

No Comments

No Known Activations