INDEX

Explanations

references to different levels of sports divisions, particularly within the NCAA

oai_token-act-pair · gpt-3.5-turbo

mentions of specific NCAA divisions

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

FORE

-0.79

pez

-0.65

packing

-0.64

isks

-0.64

aughs

-0.63

rop

-0.62

paces

-0.61

uggage

-0.61

 unaccompanied

-0.60

uated

-0.59

POSITIVE LOGITS

al

1.03

ISION

0.80

naire

0.80

alist

0.79

alis

0.78

ally

0.76

alities

0.74

ality

0.74

 Division

0.71

aler

0.71

Activations Density 0.016%

references to different levels of sports divisions, particularly within the NCAA

mentions of specific NCAA divisions

No Comments

No Known Activations