INDEX

Explanations

phrases related to the concept of attachment or connection

oai_token-act-pair · gpt-3.5-turbo

instances of the word "attached" 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 @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

schild

-0.71

 understatement

-0.70

mberg

-0.69

icist

-0.63

ouston

-0.63

MQ

-0.63

ause

-0.62

utenberg

-0.62

uddin

-0.61

 msec

-0.60

POSITIVE LOGITS

 thereto

1.44

 onto

1.04

 securely

0.90

to

0.87

 Thumbnails

0.80

 dots

0.76

 attached

0.74

 attaches

0.74

 electrodes

0.74

 wires

0.73

Activations Density 0.061%

phrases related to the concept of attachment or connection

instances of the word "attached" and its variations

No Comments

No Known Activations