INDEX

Explanations

phrases related to contrasting or opposing ideas

oai_token-act-pair · gpt-3.5-turbo

conjunctions and transitional phrases indicating relationships and flow between ideas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

uca

-0.81

etheus

-0.72

ens

-0.70

 Logged

-0.69

uits

-0.68

 SHARES

-0.68

helm

-0.68

 attRot

-0.68

,''

-0.67

ãĤ¡

-0.67

POSITIVE LOGITS

 whereas

0.83

 Conversely

0.71

 Britann

0.68

 Whereas

0.67

Medium

0.65

 others

0.62

abase

0.60

 programmers

0.59

 Colo

0.57

 Others

0.57

Activations Density 0.627%

phrases related to contrasting or opposing ideas

conjunctions and transitional phrases indicating relationships and flow between ideas

No Comments

No Known Activations