INDEX

Explanations

comparisons using the word "similar to."

oai_token-act-pair · gpt-3.5-turbo

structures and concepts that are compared or likened to each other

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

Added

-0.76

EMENT

-0.71

Limited

-0.67

erer

-0.64

whe

-0.64

Bern

-0.63

essor

-0.63

erity

-0.61

azz

-0.61

ALE

-0.61

POSITIVE LOGITS

 ours

1.01

lihood

0.89

 theirs

0.77

osponsors

0.71

oxide

0.70

anus

0.70

otin

0.67

 hers

0.67

soDeliveryDate

0.66

ptions

0.66

Activations Density 0.085%

comparisons using the word "similar to."

structures and concepts that are compared or likened to each other

No Comments

No Known Activations