INDEX

Explanations

information related to individuals and their actions in various contexts

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the introduction or recognition of individuals or entities in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 dstg

-0.76

wayne

-0.76

taboola

-0.68

.--

-0.66

%.

-0.66

aminer

-0.65

inement

-0.64

abouts

-0.63

onga

-0.62

dayName

-0.62

POSITIVE LOGITS

 prominence

0.81

 notoriety

0.78

 recently

0.74

 decom

0.68

 fame

0.68

 shortly

0.65

 Challenger

0.62

undy

0.60

 earlier

0.59

 retirement

0.58

Activations Density 0.732%

information related to individuals and their actions in various contexts

phrases indicating the introduction or recognition of individuals or entities in various contexts

No Comments

No Known Activations