INDEX

Explanations

phrases indicating a direction or inclination

oai_token-act-pair · gpt-3.5-turbo

phrases related to directional shifts or leanings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

nces

-0.74

listed

-0.72

unte

-0.70

 Lamb

-0.65

enery

-0.64

cooked

-0.64

©¶æ¥µ

-0.64

bath

-0.62

ydia

-0.62

NJ

-0.61

POSITIVE LOGITS

 toward

1.02

 favoring

0.97

 direction

0.91

 directional

0.91

 towards

0.91

 downward

0.88

 downwards

0.84

 tilt

0.84

wards

0.81

 focus

0.78

Activations Density 0.251%

phrases indicating a direction or inclination

phrases related to directional shifts or leanings

No Comments

No Known Activations