INDEX

Explanations

proper nouns and names of places or people

oai_token-act-pair · gpt-3.5-turbo

words associated with recognition or reputation

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 liability

-0.67

 track

-0.67

-0.62

 bear

-0.61

 isol

-0.61

 Buccaneers

-0.60

 sideline

-0.60

 choice

-0.60

 mitochondrial

-0.59

 grab

-0.59

POSITIVE LOGITS

aurus

0.93

imus

0.88

arten

0.80

igans

0.79

uthor

0.79

ndra

0.78

eus

0.77

ODY

0.74

urnal

0.73

ospace

0.72

Activations Density 0.340%

proper nouns and names of places or people

words associated with recognition or reputation

No Comments

No Known Activations