INDEX

Explanations

the word "mere" followed by a number

oai_token-act-pair · gpt-3.5-turbo

instances of the word "mere" indicating the trivialization or minimization of a concept or situation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

anwhile

-0.75

lain

-0.75

 Remastered

-0.70

 Lawyers

-0.68

 vigorously

-0.67

ocker

-0.66

 Nurs

-0.65

 clashed

-0.62

 admired

-0.62

 Patriarch

-0.61

POSITIVE LOGITS

 mortals

0.97

 mere

0.80

reet

0.79

 guiActiveUn

0.77

ãĥ¬

0.74

tons

0.74

ton

0.72

ername

0.69

 token

0.69

eteenth

0.68

Activations Density 0.007%

the word "mere" followed by a number

instances of the word "mere" indicating the trivialization or minimization of a concept or situation

No Comments

No Known Activations