INDEX

Explanations

proper nouns related to names or titles

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names and entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

holders

-1.01

holder

-0.79

erness

-0.73

uve

-0.72

mble

-0.68

itures

-0.64

hetically

-0.62

igating

-0.61

hetical

-0.61

 enrichment

-0.60

POSITIVE LOGITS

oint

1.06

unction

1.04

ealous

1.03

ournals

1.01

upiter

0.98

utsu

0.93

ansen

0.89

igsaw

0.89

acket

0.85

ordan

0.84

Activations Density 0.100%

proper nouns related to names or titles

proper nouns, specifically names and entities

No Comments

No Known Activations