INDEX

Explanations

instances of the word "asm" followed by a number, with higher numbers receiving stronger activations

oai_token-act-pair · gpt-3.5-turbo

words associated with various forms of "asm."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tops

-0.74

©¶æ

-0.72

llo

-0.67

WOOD

-0.64

Ey

-0.62

 Gaia

-0.61

 deduct

-0.61

VICE

-0.59

 tips

-0.59

Quantity

-0.59

POSITIVE LOGITS

ussen

1.07

olit

1.01

asm

0.99

otics

0.99

oph

0.97

iasm

0.95

atic

0.94

ophon

0.90

icity

0.90

ogyn

0.90

Activations Density 0.008%

instances of the word "asm" followed by a number, with higher numbers receiving stronger activations

words associated with various forms of "asm."

No Comments

No Known Activations