INDEX

Explanations

the word "character" with various suffixes and word formations

oai_token-act-pair · gpt-3.5-turbo

references to characters or character-driven narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

hyde

-0.74

bial

-0.68

ISSION

-0.62

 Stadium

-0.61

 Sisters

-0.61

 Spartan

-0.61

PORT

-0.60

ascus

-0.59

Ô

-0.58

 Ames

-0.58

POSITIVE LOGITS

icter

0.99

yll

0.85

uttering

0.83

ulp

0.81

etr

0.80

icol

0.80

ruct

0.79

erb

0.77

iddling

0.73

ret

0.73

Activations Density 0.034%

the word "character" with various suffixes and word formations

references to characters or character-driven narratives

No Comments

No Known Activations