INDEX

Explanations

terms related to gaining something or achieving an advantage

oai_token-act-pair · gpt-3.5-turbo

instances of the word "gain."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

senal

-0.66

 molecule

-0.64

uality

-0.62

iso

-0.61

 crying

-0.59

atis

-0.59

 Drugs

-0.57

ä¸ī

-0.57

psc

-0.56

hov

-0.55

POSITIVE LOGITS

esville

1.07

/+

0.88

 notoriety

0.85

esses

0.81

 traction

0.77

eful

0.75

icult

0.75

leon

0.75

ful

0.73

ulton

0.72

Activations Density 0.028%

terms related to gaining something or achieving an advantage

instances of the word "gain."

No Comments

No Known Activations