INDEX

Explanations

phrases mentioning personal connections or actions

oai_token-act-pair · gpt-3.5-turbo

the word "personally" and its variations, indicating a focus on personal viewpoints or experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Definitions

-0.74

eland

-0.70

 Emin

-0.69

period

-0.66

 Ends

-0.66

 Surveillance

-0.65

LY

-0.65

Dispatch

-0.65

 Stall

-0.64

 Movement

-0.63

POSITIVE LOGITS

 identifiable

1.13

 benefited

0.94

 invested

0.90

vou

0.89

 apologized

0.87

 insulted

0.86

 thanked

0.86

 offended

0.85

 intervened

0.85

ised

0.85

Activations Density 0.017%

phrases mentioning personal connections or actions

the word "personally" and its variations, indicating a focus on personal viewpoints or experiences

No Comments

No Known Activations