INDEX

Explanations

financial or administrative terms like abbreviations related to business or taxes

oai_token-act-pair · gpt-3.5-turbo

references to "PT" which likely stands for "post" or related timing indicators in a schedule

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

hold

-0.96

flies

-0.73

keepers

-0.71

wind

-0.71

ãĥ£

-0.68

plant

-0.66

ptive

-0.66

é¾įåĸļå£«

-0.66

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.65

 perspect

-0.64

POSITIVE LOGITS

ypes

1.19

EMBER

1.02

ember

1.01

icket

0.90

itious

0.87

ractor

0.87

uple

0.86

IFIED

0.83

icut

0.82

itude

0.81

Activations Density 0.004%

financial or administrative terms like abbreviations related to business or taxes

references to "PT" which likely stands for "post" or related timing indicators in a schedule

No Comments

No Known Activations