INDEX

Explanations

words related to control or suppression of something, especially through force or influence

oai_token-act-pair · gpt-3.5-turbo

words related to controlling or suppressing situations or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

optional

-0.72

Ranked

-0.70

ammy

-0.69

 placed

-0.64

finished

-0.62

ead

-0.62

ever

-0.61

umbered

-0.60

args

-0.60

isd

-0.59

POSITIVE LOGITS

 havoc

0.80

ieri

0.77

 Flavoring

0.74

 irritation

0.71

 airflow

0.68

¿½

0.66

 fury

0.66

 clouds

0.65

 hostilities

0.64

ILA

0.64

Activations Density 0.097%

words related to control or suppression of something, especially through force or influence

words related to controlling or suppressing situations or entities

No Comments

No Known Activations