INDEX

Explanations

words related to events, sequences, or actions that have already happened

oai_token-act-pair · gpt-3.5-turbo

instances of the word "ced" and variations of it

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 motto

-0.68

Balt

-0.67

 Scotia

-0.67

 contrace

-0.66

ISH

-0.66

 tremend

-0.64

 Bravo

-0.63

 Gaga

-0.63

isha

-0.63

 Barkley

-0.62

POSITIVE LOGITS

ragon

0.93

dit

0.91

ez

0.86

ures

0.84

ornia

0.84

0.83

gments

0.81

umbnail

0.81

enced

0.80

ety

0.80

Activations Density 0.018%

words related to events, sequences, or actions that have already happened

instances of the word "ced" and variations of it

No Comments

No Known Activations