INDEX

Explanations

phrases indicating someone being caught in a negative or suspicious situation

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 pronoun

-0.74

eth

-0.74

 pronouns

-0.71

ciplinary

-0.70

 helicop

-0.70

htar

-0.70

xual

-0.68

umably

-0.68

estate

-0.66

ensible

-0.66

POSITIVE LOGITS

ãĤµ

0.85

ãĤ¤

0.76

 Tracks

0.74

 Luffy

0.74

Rai

0.73

ãĥĩãĤ£

0.72

ãĤ¤ãĥĪ

0.72

 caught

0.72

ãĥĥãĥī

0.71

Frames

0.70

Activations Density 0.012%

phrases indicating someone being caught in a negative or suspicious situation

No Comments

No Known Activations

phrases indicating someone being caught in a negative or suspicious situation

No Comments

No Known Activations