INDEX

Explanations

phrases related to actions of turning or handing over

oai_token-act-pair · gpt-3.5-turbo

variations of the word "turn."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

è¦ļéĨĴ

-0.88

ropolitan

-0.68

ording

-0.67

mma

-0.62

bour

-0.60

ordan

-0.59

assment

-0.59

lain

-0.59

ority

-0.59

capacity

-0.59

POSITIVE LOGITS

 into

0.83

Ī

0.77

awa

0.77

 traitor

0.74

coat

0.74

 Into

0.72

¸

0.71

 informant

0.71

 INTO

0.70

buck

0.70

Activations Density 0.045%

phrases related to actions of turning or handing over

variations of the word "turn."

No Comments

No Known Activations