INDEX

Explanations

references to a specific term or acronym "CR"

oai_token-act-pair · gpt-3.5-turbo

references to "CR" which likely indicates a specific term or acronym in context, related to various topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Ko

-0.77

ften

-0.75

oat

-0.73

uta

-0.71

 Mata

-0.69

apy

-0.66

Uk

-0.66

 Kent

-0.66

oslav

-0.65

noon

-0.64

POSITIVE LOGITS

CR

3.65

CR

2.55

CRA

1.49

CL

1.41

FR

1.40

DR

1.38

Cr

1.22

cr

1.21

TR

1.21

CRE

1.21

Activations Density 0.016%

references to a specific term or acronym "CR"

references to "CR" which likely indicates a specific term or acronym in context, related to various topics

No Comments

No Known Activations