INDEX

Explanations

phrases related to approving or authorizing something

oai_token-act-pair · gpt-3.5-turbo

terms related to approval or authorization

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

fact

-0.65

alus

-0.64

TON

-0.64

ellow

-0.63

ngth

-0.63

anwhile

-0.63

EMS

-0.62

orld

-0.62

raltar

-0.61

irrel

-0.61

POSITIVE LOGITS

 unanimously

0.94

 authorizing

0.91

 waivers

0.88

 approval

0.86

 approvals

0.82

 approving

0.82

 petitions

0.75

 approves

0.72

 warrants

0.71

 approve

0.69

Activations Density 0.067%

phrases related to approving or authorizing something

terms related to approval or authorization

No Comments

No Known Activations