INDEX

Explanations

statements or assertions

oai_token-act-pair · gpt-3.5-turbo

expressions indicating thoughts or feelings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Whilst

-1.05

Âł

-0.90

wik

-0.87

 Â®

-0.81

 whilst

-0.80

 Citiz

-0.78

',

-0.77

XY

-0.77

âĢĲ

-0.76

".[

-0.76

POSITIVE LOGITS

 teammates

1.20

 coach

1.18

 postseason

1.15

 Belichick

1.11

 offseason

1.11

 teammate

1.09

 quarterback

1.06

 coaches

1.06

 Manziel

1.06

 preseason

1.04

Activations Density 0.782%

statements or assertions

expressions indicating thoughts or feelings

No Comments

No Known Activations