INDEX

Explanations

proper nouns, specifically those related to organizations or places

oai_token-act-pair · gpt-3.5-turbo

the abbreviation "MI" frequently associated with various contexts or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

size

-0.76

ding

-0.75

language

-0.73

ional

-0.73

sworth

-0.73

served

-0.70

holder

-0.70

sword

-0.70

mans

-0.70

holders

-0.69

POSITIVE LOGITS

AMI

1.11

HAEL

1.01

PS

0.98

uppet

0.95

ovember

0.85

KE

0.81

MI

0.79

XT

0.79

OSH

0.78

eredith

0.76

Activations Density 0.007%

proper nouns, specifically those related to organizations or places

the abbreviation "MI" frequently associated with various contexts or entities

No Comments

No Known Activations