INDEX

Explanations

phrases indicating conflict or competition

oai_token-act-pair · gpt-3.5-turbo

ellipsis or unfinished thoughts in sentences

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

icum

-0.85

odi

-0.73

icka

-0.72

oit

-0.70

 condem

-0.70

enei

-0.70

udding

-0.69

lings

-0.69

terday

-0.68

ivers

-0.68

POSITIVE LOGITS

BUT

1.02

 Written

0.88

 Appears

0.82

 Continue

0.77

 Continued

0.75

wait

0.74

 Wait

0.72

âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦

0.71

Goo

0.71

 âĢİ

0.70

Activations Density 0.013%

phrases indicating conflict or competition

ellipsis or unfinished thoughts in sentences

No Comments

No Known Activations