INDEX

Explanations

negative statements or situations

oai_token-act-pair · gpt-3.5-turbo

phrases containing the word "not" indicating negation or rejection

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Spotlight

-0.77

èĢħ

-0.68

 Advice

-0.68

ãĥ¼ãĥĨãĤ£

-0.67

 Inventory

-0.66

former

-0.63

weekly

-0.62

Site

-0.62

Reviewer

-0.62

 Tycoon

-0.61

POSITIVE LOGITS

icably

1.13

icable

1.02

 bother

0.97

epad

0.93

 necessarily

0.92

 exist

0.91

ifying

0.82

 adequately

0.80

hin

0.79

 existed

0.77

Activations Density 0.224%

negative statements or situations

phrases containing the word "not" indicating negation or rejection

No Comments

No Known Activations