INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

occurrences of specific letters or letter combinations, focusing on their frequency

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ulhu

-0.75

pace

-0.72

wcs

-0.68

akeru

-0.62

DragonMagazine

-0.61

iolet

-0.59

swick

-0.58

å§«

-0.57

intosh

-0.56

nown

-0.56

POSITIVE LOGITS

uala

0.74

ueller

0.61

renheit

0.61

Zak

0.55

enthal

0.53

ilty

0.52

 Vault

0.52

FW

0.51

 Sharif

0.51

Grand

0.50

Activations Density 0.283%

names of individuals

occurrences of specific letters or letter combinations, focusing on their frequency

No Comments

No Known Activations