INDEX

Explanations

numerical data presented with a comparison or distinction

oai_token-act-pair · gpt-3.5-turbo

the word "respectively" and its variations, indicating comparisons or correspondences between items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

bern

-0.78

ger

-0.75

rak

-0.74

iens

-0.72

bor

-0.70

 Cosponsors

-0.68

 Defenders

-0.68

gravity

-0.67

lo

-0.66

wordpress

-0.65

POSITIVE LOGITS

 guiActiveUn

1.03

 exha

0.84

 eleph

0.82

 sexes

0.79

 weep

0.79

 detach

0.78

 redund

0.73

 situated

0.73

 unus

0.73

 halves

0.72

Activations Density 0.005%

numerical data presented with a comparison or distinction

the word "respectively" and its variations, indicating comparisons or correspondences between items

No Comments

No Known Activations