INDEX

Explanations

words related to receiving some type of recognition or honor

oai_token-act-pair · gpt-3.5-turbo

references to awards 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

loo

-0.69

ths

-0.68

 Occupations

-0.68

\/\/

-0.67

Klu

-0.67

 Pastebin

-0.66

 Sins

-0.65

STD

-0.64

ãĥ¯ãĥ³

-0.63

ullivan

-0.62

POSITIVE LOGITS

 awards

1.15

 award

1.07

 laure

1.05

 awarded

0.99

 prizes

0.93

 awarding

0.91

Winner

0.90

 winner

0.89

 prize

0.89

winning

0.87

Activations Density 0.013%

words related to receiving some type of recognition or honor

references to awards and recognition

No Comments

No Known Activations