INDEX

Explanations

verbs related to negative outcomes or consequences, particularly the word "lose."

oai_token-act-pair · gpt-3.5-turbo

the term "lose" in various contexts

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

 plun

-0.64

orig

-0.63

 Austral

-0.59

osi

-0.59

 Municip

-0.58

GROUP

-0.58

 primed

-0.58

opot

-0.57

 Janeiro

-0.57

ATT

-0.57

POSITIVE LOGITS

iem

0.81

itaire

0.79

esty

0.79

luster

0.78

 aversion

0.77

itive

0.76

veyard

0.75

 virginity

0.71

esses

0.71

soever

0.71

Activations Density 0.028%

verbs related to negative outcomes or consequences, particularly the word "lose."

the term "lose" in various contexts

No Comments

No Known Activations