INDEX

Explanations

mentions of writers

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "writer" 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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ADRA

-0.79

rals

-0.77

illon

-0.76

 Lumpur

-0.73

aband

-0.72

ĸļ

-0.71

ypes

-0.70

ibaba

-0.70

eneg

-0.70

asonic

-0.68

POSITIVE LOGITS

 writer

0.95

writer

0.94

 laureate

0.91

uscript

0.87

 fiction

0.86

writing

0.85

writ

0.82

haw

0.81

 Beware

0.78

itatively

0.76

Activations Density 0.028%

mentions of writers

occurrences of the word "writer" in various contexts

No Comments

No Known Activations