INDEX

Explanations

social theories or common beliefs

oai_token-act-pair · gpt-3.5-turbo

references to commonly held beliefs or theories

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

ILE

-0.84

Interstitial

-0.82

andestine

-0.76

phabet

-0.75

CLASSIFIED

-0.73

quished

-0.70

ktop

-0.67

 Anniversary

-0.64

 Doors

-0.64

Bonus

-0.63

POSITIVE LOGITS

 persists

0.97

 prevailed

0.96

pos

0.95

 blames

0.95

 belief

0.91

 asserts

0.91

 assumption

0.90

 errone

0.90

 assumes

0.87

 perpet

0.86

Activations Density 0.270%

social theories or common beliefs

references to commonly held beliefs or theories

No Comments

No Known Activations