INDEX

Explanations

lists or enumerations with clear indicators like "Here are two of them."

oai_token-act-pair · gpt-3.5-turbo

phrases indicating lists or groupings of items

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 commodities

-0.75

Financial

-0.74

 counselling

-0.74

 fostering

-0.74

angering

-0.72

 administrations

-0.71

 stake

-0.71

 deterrence

-0.69

 counseling

-0.69

etermined

-0.69

POSITIVE LOGITS

 gorgeous

0.88

 pics

0.87

 Beaut

0.86

 prett

0.84

itled

0.83

 Delicious

0.83

 beautifully

0.82

 funn

0.80

unny

0.80

ozy

0.79

Activations Density 0.846%

lists or enumerations with clear indicators like "Here are two of them."

phrases indicating lists or groupings of items

No Comments

No Known Activations