INDEX

Explanations

mentions of statistics or comparison of parts to the whole

oai_token-act-pair · gpt-3.5-turbo

phrases that refer to collective entities or groups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

andon

-0.73

orange

-0.66

bots

-0.62

iron

-0.62

antine

-0.61

Ru

-0.60

ATURES

-0.60

 Arri

-0.59

Roy

-0.59

doms

-0.59

POSITIVE LOGITS

 whole

1.59

 result

1.53

 consequence

1.33

 percentage

1.04

 matter

0.95

 cohesive

0.95

 totality

0.94

 standalone

0.90

 Result

0.84

 bloc

0.83

Activations Density 0.068%

mentions of statistics or comparison of parts to the whole

phrases that refer to collective entities or groups

No Comments

No Known Activations