INDEX

Explanations

The main thing this neuron does is find statements expressing opinions or beliefs

oai_token-act-pair · gpt-3.5-turbo

expressions of personal identity and self-reference

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

hess

-0.72

 sidx

-0.68

udeau

-0.64

atives

-0.64

racuse

-0.64

eline

-0.63

estones

-0.63

Scroll

-0.61

akings

-0.61

rival

-0.61

POSITIVE LOGITS

 merely

1.46

 simply

1.25

not

1.08

 purely

1.05

 instead

1.05

 just

0.98

 actually

0.97

 neither

0.95

 certainly

0.92

NOT

0.91

Activations Density 0.241%

The main thing this neuron does is find statements expressing opinions or beliefs

expressions of personal identity and self-reference

No Comments

No Known Activations