INDEX

Explanations

phrases related to behind-the-scenes activities

oai_token-act-pair · gpt-3.5-turbo

phrases indicating behind-the-scenes activities or contexts

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

theless

-0.79

 Rica

-0.70

awaru

-0.70

ertodd

-0.69

å§«

-0.66

DragonMagazine

-0.66

 behavi

-0.64

ADRA

-0.62

 Debor

-0.61

udo

-0.61

POSITIVE LOGITS

same

0.95

purpose

0.88

door

0.84

prem

0.84

best

0.83

est

0.83

middle

0.81

advertising

0.81

committee

0.79

grounds

0.78

Activations Density 0.029%

phrases related to behind-the-scenes activities

phrases indicating behind-the-scenes activities or contexts

No Comments

No Known Activations