INDEX

Explanations

phrases related to relative position or status

oai_token-act-pair · gpt-3.5-turbo

phrases that signify being out of context or not aligned with expectations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

nai

-0.78

advertisement

-0.75

incial

-0.75

utical

-0.70

ighth

-0.68

arnaev

-0.64

è£ħ

-0.63

ijing

-0.62

Ń·

-0.62

shown

-0.61

POSITIVE LOGITS

 altogether

0.78

 boredom

0.69

 mouths

0.67

 entirely

0.67

alus

0.66

vae

0.65

 toile

0.65

ville

0.64

 frying

0.61

 drawer

0.60

Activations Density 0.110%

phrases related to relative position or status

phrases that signify being out of context or not aligned with expectations

No Comments

No Known Activations