INDEX

Explanations

adjectives describing different levels or types of actions or behaviors

oai_token-act-pair · gpt-3.5-turbo

adjectives and phrases indicating levels of appropriateness or measures being taken

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

-+-+

-0.80

arthed

-0.73

ntil

-0.72

Unt

-0.71

Ü

-0.68

IOR

-0.65

bably

-0.64

uce

-0.61

tions

-0.61

eret

-0.61

POSITIVE LOGITS

 steps

1.30

 precautions

1.19

 liberties

1.09

 step

1.05

 risks

1.05

 oath

1.05

 strides

1.04

steps

1.03

 precaution

0.99

 Steps

0.98

Activations Density 0.101%

adjectives describing different levels or types of actions or behaviors

adjectives and phrases indicating levels of appropriateness or measures being taken

No Comments

No Known Activations