INDEX

Explanations

the word "sum" along with numerical values

oai_token-act-pair · gpt-3.5-turbo

references to cumulative totals or sums

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

 livest

-0.70

 Defenders

-0.70

 surn

-0.69

 Ferry

-0.68

 Kers

-0.66

 foreclosure

-0.66

eele

-0.62

 blinking

-0.57

 Centers

-0.57

 Brewers

-0.57

POSITIVE LOGITS

ming

1.37

mits

1.07

mers

1.03

sum

0.97

otaur

0.92

ulative

0.91

atra

0.87

sum

0.86

iture

0.82

mes

0.81

Activations Density 0.008%

the word "sum" along with numerical values

references to cumulative totals or sums

No Comments

No Known Activations