INDEX

Explanations

The neuron activates for the word "of"

oai_token-act-pair · gpt-3.5-turbo

the phrase "Of course."

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

yles

-0.64

ocene

-0.63

 cleaners

-0.62

pieces

-0.62

Ô

-0.61

 reperto

-0.61

dot

-0.61

"}],"

-0.60

è¦ļéĨĴ

-0.59

sew

-0.58

POSITIVE LOGITS

 course

1.34

ortunately

0.99

course

0.98

elia

0.93

 Course

0.89

sted

0.87

 sorts

0.86

onso

0.81

icial

0.77

rontal

0.77

Activations Density 0.029%

The neuron activates for the word "of"

the phrase "Of course."

No Comments

No Known Activations