INDEX

Explanations

mention of the word "gaming" with varying activations

oai_token-act-pair · gpt-3.5-turbo

references to gaming

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

eous

-0.74

tain

-0.70

taining

-0.65

drawn

-0.64

iflower

-0.63

von

-0.63

headed

-0.62

Jol

-0.62

inators

-0.62

ioxide

-0.61

POSITIVE LOGITS

 consoles

0.91

 gaming

0.90

manship

0.81

cube

0.81

 gamers

0.79

 peripher

0.76

 prowess

0.75

HAM

0.75

 videog

0.74

 cafes

0.73

Activations Density 0.014%

mention of the word "gaming" with varying activations

references to gaming

No Comments

No Known Activations