INDEX

Explanations

the word "Most" followed by various contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "most" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

rompt

-0.74

 Travels

-0.68

DOT

-0.67

arium

-0.67

 CARD

-0.66

thur

-0.66

pload

-0.66

pless

-0.63

LOT

-0.60

plan

-0.59

POSITIVE LOGITS

 importantly

1.01

 Wanted

0.95

 Helpful

0.90

 Likely

0.90

 Important

0.88

 important

0.82

entimes

0.81

likely

0.80

 notable

0.80

egu

0.75

Activations Density 0.048%

the word "Most" followed by various contexts

instances of the word "most" in various contexts

No Comments

No Known Activations