INDEX

Explanations

words related to public appearances or actions in the media

oai_token-act-pair · gpt-3.5-turbo

instances of the word "appear" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Haram

-0.70

 Mahjong

-0.65

warts

-0.64

dp

-0.62

NRS

-0.61

 Habit

-0.60

ISO

-0.58

iott

-0.58

fail

-0.57

POSITIVE LOGITS

 onstage

1.07

 alongside

0.99

 briefly

0.86

on

0.83

 before

0.82

 prominently

0.80

 beside

0.79

united

0.76

 nude

0.75

 opposite

0.74

Activations Density 0.063%

words related to public appearances or actions in the media

instances of the word "appear" in various contexts

No Comments

No Known Activations