INDEX

Explanations

words related to honors, awards, and recognition

oai_token-act-pair · gpt-3.5-turbo

terms related to academic honors and recognition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

arre

-0.73

okin

-0.72

acco

-0.68

 DOWN

-0.60

iHUD

-0.60

ulhu

-0.60

 Reloaded

-0.59

Hyd

-0.59

 Kindle

-0.59

SE

-0.59

POSITIVE LOGITS

 honoured

1.08

 honors

1.06

brance

0.95

 acknowled

0.91

abilia

0.91

 honour

0.88

 bestowed

0.87

inguished

0.85

 honored

0.83

 Lauder

0.82

Activations Density 0.073%

words related to honors, awards, and recognition

terms related to academic honors and recognition

No Comments

No Known Activations