INDEX

Explanations

actions involving physical interaction with objects or people

oai_token-act-pair · gpt-3.5-turbo

articles preceding descriptions of physical actions or contexts

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

ACTION

-0.82

OTOS

-0.77

 Accounts

-0.75

Events

-0.74

effective

-0.74

Hours

-0.74

liest

-0.73

 assessments

-0.73

apps

-0.72

 Cases

-0.71

POSITIVE LOGITS

 doorway

1.04

 pile

1.02

 bunch

1.01

 cliff

1.01

 pedest

0.98

 ledge

0.97

 boulder

0.96

 desk

0.95

 hill

0.94

 mound

0.93

Activations Density 0.226%

actions involving physical interaction with objects or people

articles preceding descriptions of physical actions or contexts

No Comments

No Known Activations