INDEX

Explanations

the word "Perhaps"

oai_token-act-pair · gpt-3.5-turbo

statements that express uncertainty or speculation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

iosis

-0.73

atches

-0.71

 blender

-0.70

emy

-0.69

ords

-0.68

arp

-0.66

rogens

-0.66

lined

-0.65

ursed

-0.64

ocaust

-0.64

POSITIVE LOGITS

Perhaps

0.80

 "$:/

0.78

 Perhaps

0.76

 unsurprisingly

0.75

theless

0.73

perhaps

0.71

 sensing

0.70

 tempted

0.69

Comment

0.68

Discuss

0.68

Activations Density 0.009%

the word "Perhaps"

statements that express uncertainty or speculation

No Comments

No Known Activations