INDEX

Explanations

words related to different viewpoints or perspectives

oai_token-act-pair · gpt-3.5-turbo

phrases related to differing viewpoints or angles of analysis

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

arus

-0.82

ãĥīãĥ©

-0.79

ãģį

-0.75

usters

-0.70

 notor

-0.68

redo

-0.68

pload

-0.66

ãĥĪ

-0.66

planet

-0.63

bart

-0.63

POSITIVE LOGITS

 standpoint

0.84

 onwards

0.83

 alone

0.81

 alike

0.76

 anyway

0.73

 perspective

0.72

($

0.71

0.70

 onward

0.70

enance

0.69

Activations Density 0.165%

words related to different viewpoints or perspectives

phrases related to differing viewpoints or angles of analysis

No Comments

No Known Activations