INDEX

Explanations

words related to letters, like the alphabet character

oai_token-act-pair · gpt-3.5-turbo

references to letters and written characters

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

rolet

-0.88

oleon

-0.77

amera

-0.76

ESCO

-0.69

Govern

-0.68

Latest

-0.67

 WATCH

-0.67

OURCE

-0.65

allows

-0.65

Exec

-0.63

POSITIVE LOGITS

 letters

1.00

letter

0.90

sworth

0.84

Letter

0.82

mith

0.81

 letter

0.80

worms

0.79

ymes

0.78

 lett

0.76

 Cipher

0.76

Activations Density 0.013%

words related to letters, like the alphabet character

references to letters and written characters

No Comments

No Known Activations