INDEX

Explanations

proper nouns referring to awards, agreements, or significant historical events

oai_token-act-pair · gpt-3.5-turbo

terms related to significant entities, honors, and formal agreements

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

iable

-0.81

etheless

-0.76

theless

-0.75

actionDate

-0.70

cedented

-0.69

posted

-0.69

 manag

-0.68

 directions

-0.68

fitting

-0.68

dated

-0.68

POSITIVE LOGITS

 Collection

1.05

 Initiative

1.03

 Cooperative

0.97

 Consortium

0.92

 Prize

0.91

 Project

0.91

 Massacre

0.90

 Circuit

0.90

 Orchestra

0.88

 Foundation

0.88

Activations Density 0.265%

proper nouns referring to awards, agreements, or significant historical events

terms related to significant entities, honors, and formal agreements

No Comments

No Known Activations