INDEX

Explanations

phrases indicating speaking or providing information under a condition of anonymity

oai_token-act-pair · gpt-3.5-turbo

phrases referencing speaking or communication in a context of anonymity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ratulations

-0.73

holes

-0.71

naire

-0.69

BIP

-0.69

edia

-0.66

aurus

-0.65

$$

-0.64

otle

-0.64

Ã©n

-0.63

Poll

-0.63

POSITIVE LOGITS

 behalf

1.65

 condition

1.25

 sidelines

1.01

 occasion

0.97

 Skype

0.90

 phone

0.88

 anonymity

0.85

etime

0.84

 camera

0.83

 background

0.82

Activations Density 0.105%

phrases indicating speaking or providing information under a condition of anonymity

phrases referencing speaking or communication in a context of anonymity

No Comments

No Known Activations