INDEX

Explanations

negative situations or actions

oai_token-act-pair · gpt-3.5-turbo

punctuations and symbols indicating section breaks or end of thought

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

 slump

-0.72

ĪĴ

-0.68

 exhib

-0.67

 undermin

-0.67

 princ

-0.66

 tenancy

-0.66

 isot

-0.65

 administ

-0.65

 exha

-0.65

 exhibitions

-0.65

POSITIVE LOGITS

<|endoftext|>

1.30

↵↵

1.19

↵

1.13

 ********************************

1.07

 Anyway

0.99

 âĶľâĶĢâĶĢ

0.94

 Originally

0.92

[/

0.91

↵Âł

0.89

||||

0.89

Activations Density 0.113%

negative situations or actions

punctuations and symbols indicating section breaks or end of thought

No Comments

No Known Activations