INDEX

Explanations

proper nouns and verbs related to achievements or awards

oai_token-act-pair · gpt-3.5-turbo

mentions of awards and accolades

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

aturdays

-0.71

ivo

-0.69

sin

-0.65

town

-0.65

rians

-0.65

 flows

-0.65

erers

-0.64

 Sorce

-0.64

bleacher

-0.64

markets

-0.63

POSITIVE LOGITS

 accol

1.10

 Nobel

1.09

 prestigious

1.07

 medals

1.02

 Pulitzer

0.99

 awards

0.98

 prizes

0.98

 honorary

0.97

 nomination

0.95

Nob

0.93

Activations Density 0.183%

proper nouns and verbs related to achievements or awards

mentions of awards and accolades

No Comments

No Known Activations