INDEX

Explanations

proper nouns related to businesses and people

oai_token-act-pair · gpt-3.5-turbo

references to specific brand names or organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

CAST

-0.85

Temperature

-0.73

sheet

-0.73

OME

-0.70

ÃįÃį

-0.69

dies

-0.68

LLOW

-0.67

chool

-0.67

Values

-0.66

bleacher

-0.66

POSITIVE LOGITS

ovic

1.02

agement

0.97

jar

0.96

ufact

0.91

ipal

0.89

agements

0.88

amiya

0.86

agos

0.85

aneous

0.85

ova

0.84

Activations Density 0.010%

proper nouns related to businesses and people

references to specific brand names or organizations

No Comments

No Known Activations