INDEX

Explanations

mentions of trying to understand or interpreting desired actions

oai_token-act-pair · gpt-3.5-turbo

phrases related to understanding and communication

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Flickr

-0.81

 outraged

-0.81

 ACLU

-0.77

 purportedly

-0.75

 triv

-0.73

mint

-0.71

uphem

-0.70

 conveniently

-0.70

 Slate

-0.70

 doubtless

-0.70

POSITIVE LOGITS

 coaching

1.38

 player

1.11

 coaches

1.10

Players

1.09

 talent

1.06

 Coach

1.06

Coach

1.06

 players

1.05

 teammates

1.05

 rookies

1.04

Activations Density 0.844%

mentions of trying to understand or interpreting desired actions

phrases related to understanding and communication

No Comments

No Known Activations