INDEX

Explanations

The neuron is activated by sentences indicating the beginning or end of an article or section

oai_token-act-pair · gpt-3.5-turbo

sentences emphasizing the concept of truth or factual statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

peg

-0.61

uese

-0.60

inav

-0.59

 Telesc

-0.58

iates

-0.57

 Towers

-0.57

iture

-0.55

 encounters

-0.54

itches

-0.53

 slips

-0.53

POSITIVE LOGITS

rael

1.13

senal

0.95

nt

0.95

cussion

0.94

 hereby

0.92

 definitely

0.90

olate

0.88

olation

0.83

not

0.83

 unlikely

0.82

Activations Density 0.576%

The neuron is activated by sentences indicating the beginning or end of an article or section

sentences emphasizing the concept of truth or factual statements

No Comments

No Known Activations