INDEX

Explanations

proper nouns, particularly people's names

oai_token-act-pair · gpt-3.5-turbo

names of individuals or characters

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

ually

-0.74

#$

-0.72

forced

-0.64

lihood

-0.63

acity

-0.62

PokÃ©mon

-0.62

rophe

-0.62

 Cyborg

-0.59

 Creator

-0.58

 corridors

-0.58

POSITIVE LOGITS

mie

0.90

iners

0.90

 Seym

0.84

sie

0.83

inery

0.82

zbollah

0.80

sat

0.79

alf

0.79

ginx

0.79

ineries

0.78

Activations Density 0.042%

proper nouns, particularly people's names

names of individuals or characters

No Comments

No Known Activations