INDEX

Explanations

phrases containing the word "liar" or variations thereof.

oai_token-act-pair · gpt-3.5-turbo

references to specific names or terms associated with cultural contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Rasmussen

-0.69

CRIP

-0.65

ETF

-0.63

 Fast

-0.63

Pax

-0.60

 stances

-0.60

AX

-0.60

TD

-0.59

 Vend

-0.58

 Volunteer

-0.57

POSITIVE LOGITS

iar

1.31

amas

0.89

oga

0.87

osaurs

0.82

vier

0.81

ulence

0.79

isen

0.78

ises

0.77

osc

0.76

monary

0.76

Activations Density 0.009%

phrases containing the word "liar" or variations thereof.

references to specific names or terms associated with cultural contexts

No Comments

No Known Activations