INDEX

Explanations

instances where an action or information is pointed out or suggested

oai_token-act-pair · gpt-3.5-turbo

instances of the word "indicated."

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

 remake

-0.65

venge

-0.61

oir

-0.60

Behind

-0.60

 hairc

-0.59

haul

-0.58

js

-0.58

bies

-0.57

 scra

-0.57

 collaborations

-0.57

POSITIVE LOGITS

 indicated

3.41

 signaled

1.86

 hinted

1.77

 suggested

1.75

 stated

1.74

 indicate

1.68

 indicates

1.67

 expressed

1.52

 implied

1.51

 indicating

1.47

Activations Density 0.011%

instances where an action or information is pointed out or suggested

instances of the word "indicated."

No Comments

No Known Activations