INDEX

Explanations

words related to logical reasoning and causation

oai_token-act-pair · gpt-3.5-turbo

phrases that emphasize causation or conclusions drawn from prior statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Beaver

-0.66

 Cascade

-0.66

 Feld

-0.66

 penny

-0.66

 Omaha

-0.64

 Patty

-0.64

 Tulsa

-0.63

ja

-0.62

 Defenders

-0.60

 Berkshire

-0.59

POSITIVE LOGITS

forth

1.53

 guiActiveUn

0.80

 ineligible

0.76

 deserve

0.75

xual

0.74

 cannot

0.74

çĶŁ

0.73

uracy

0.73

 excludes

0.72

 shouldn

0.72

Activations Density 0.039%

words related to logical reasoning and causation

phrases that emphasize causation or conclusions drawn from prior statements

No Comments

No Known Activations