INDEX

Explanations

ordinal numbers indicating ranking or position

oai_token-act-pair · gpt-3.5-turbo

ordinal rankings and positions in sports contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 energies

-0.63

Rules

-0.61

 Brist

-0.60

pi

-0.60

dep

-0.59

 leases

-0.59

 RELEASE

-0.57

 raids

-0.56

iann

-0.55

 confessions

-0.55

POSITIVE LOGITS

retty

0.90

 overall

0.86

 among

0.85

BuyableInstoreAndOnline

0.85

 amongst

0.82

Tier

0.81

 percentile

0.81

 athlet

0.80

 nationally

0.76

swick

0.75

Activations Density 0.082%

ordinal numbers indicating ranking or position

ordinal rankings and positions in sports contexts

No Comments

No Known Activations