INDEX

Explanations

adjectives used to describe qualities or features of objects

oai_token-act-pair · gpt-3.5-turbo

terms related to societal structure and function

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

theless

-0.77

 embassies

-0.75

 McGee

-0.73

 iPads

-0.73

clips

-0.70

 Zombies

-0.69

videos

-0.66

 temples

-0.65

 URLs

-0.64

 dams

-0.62

POSITIVE LOGITS

ainer

0.72

 standpoint

0.71

 performer

0.71

ollower

0.69

uristic

0.69

albeit

0.67

agonist

0.66

ifier

0.65

bike

0.64

ader

0.63

Activations Density 0.501%

adjectives used to describe qualities or features of objects

terms related to societal structure and function

No Comments

No Known Activations