INDEX

Explanations

phrases comparing something to its previous state or comparing it to a standard or expectation

oai_token-act-pair · gpt-3.5-turbo

phrases that convey comparisons or changes over time

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Ward

-0.67

 ward

-0.66

 Baron

-0.61

 Kelley

-0.58

 dissenting

-0.56

 Mund

-0.56

 miscarriage

-0.55

mma

-0.55

eria

-0.53

 Mish

-0.52

POSITIVE LOGITS

 ought

0.88

 sounds

0.82

 deserves

0.79

 belongs

0.78

 resembles

0.77

 relates

0.76

 Gets

0.76

 seems

0.76

 appears

0.75

 Appears

0.75

Activations Density 0.076%

phrases comparing something to its previous state or comparing it to a standard or expectation

phrases that convey comparisons or changes over time

No Comments

No Known Activations