INDEX

Explanations

lists or sequences of items

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "list" and its variations

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

 Aber

-0.67

asio

-0.65

 sailing

-0.64

rir

-0.64

 seas

-0.64

 turf

-0.63

perty

-0.62

 pounding

-0.61

 merry

-0.60

lat

-0.60

POSITIVE LOGITS

ening

1.15

erv

1.10

ener

1.01

eners

0.97

ing

0.87

enf

0.86

uce

0.86

list

0.85

erve

0.84

enable

0.83

Activations Density 0.021%

lists or sequences of items

occurrences of the word "list" and its variations

No Comments

No Known Activations