INDEX

Explanations

phrases indicating the responsibility or decision-making authority of a person or entity

oai_token-act-pair · gpt-3.5-turbo

phrases indicating responsibility or agency, particularly the phrase "up to."

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

 Advertisement

-0.74

multipl

-0.70

afety

-0.62

 promo

-0.62

irect

-0.61

 revolves

-0.61

 referenced

-0.60

heading

-0.58

 Born

-0.58

need

-0.57

POSITIVE LOGITS

 pasture

0.88

perty

0.83

pload

0.81

orthy

0.79

riage

0.78

asty

0.76

ggles

0.72

 date

0.70

NRS

0.70

 ensure

0.69

Activations Density 0.051%

phrases indicating the responsibility or decision-making authority of a person or entity

phrases indicating responsibility or agency, particularly the phrase "up to."

No Comments

No Known Activations