INDEX

Explanations

phrases related to achievements and awards

oai_token-act-pair · gpt-3.5-turbo

references to multiple-time athletic achievements or titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enance

-0.78

dyl

-0.77

hops

-0.77

uden

-0.76

conom

-0.76

speak

-0.74

ships

-0.72

pathy

-0.71

plot

-0.71

rooms

-0.70

POSITIVE LOGITS

 winner

1.33

 champ

1.32

 champion

1.28

 Olympic

1.23

 Olymp

1.19

MVP

1.11

 recipient

1.09

 Champion

1.07

 runner

1.06

 contender

1.04

Activations Density 0.096%

phrases related to achievements and awards

references to multiple-time athletic achievements or titles

No Comments

No Known Activations