INDEX

Explanations

mentions of predecessors or counterparts

oai_token-act-pair · gpt-3.5-turbo

terms that refer to prior entities or versions associated with a current subject

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

awar

-0.71

erer

-0.69

erest

-0.67

omy

-0.65

der

-0.62

floor

-0.62

clusions

-0.60

oros

-0.60

cluded

-0.60

atari

-0.60

POSITIVE LOGITS

é¾įåĸļå£«

0.81

arios

0.64

hip

0.64

 SERV

0.63

cules

0.62

adeon

0.62

cule

0.61

reviewed

0.61

 tasked

0.61

sonian

0.61

Activations Density 0.053%

mentions of predecessors or counterparts

terms that refer to prior entities or versions associated with a current subject

No Comments

No Known Activations