INDEX

Explanations

computer-related commands or instructions

oai_token-act-pair · gpt-3.5-turbo

words related to commands and command execution

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ylan

-0.71

yr

-0.69

season

-0.68

sted

-0.64

 Scholarship

-0.62

scar

-0.61

 tourism

-0.61

oli

-0.59

igious

-0.59

osa

-0.59

POSITIVE LOGITS

 commands

3.79

 Commands

2.78

 command

2.03

 commanded

1.83

command

1.72

Command

1.65

 instructions

1.60

 Command

1.58

 orders

1.58

 directives

1.58

Activations Density 0.012%

computer-related commands or instructions

words related to commands and command execution

No Comments

No Known Activations