INDEX

Explanations

The neuron is looking for words related to being beset or plagued by something negative

oai_token-act-pair · gpt-3.5-turbo

references to the term "best."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Marijuana

-0.72

 Mariners

-0.65

 District

-0.65

 Capitals

-0.65

Sul

-0.64

 Roosevelt

-0.62

 Packers

-0.61

osaurs

-0.61

 Reds

-0.61

 Sounders

-0.59

POSITIVE LOGITS

 challeng

1.15

bes

1.12

entimes

1.00

semb

0.83

sembly

0.83

iege

0.83

poke

0.79

pse

0.78

cape

0.78

 predec

0.77

Activations Density 0.006%

The neuron is looking for words related to being beset or plagued by something negative

references to the term "best."

No Comments

No Known Activations