INDEX

Explanations

mentions of hierarchical structures or promotions within groups

oai_token-act-pair · gpt-3.5-turbo

the word "ranks" and its various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

argo

-0.88

 Robo

-0.65

 Archangel

-0.65

\/\/

-0.62

uras

-0.61

è¦ļéĨĴ

-0.61

positive

-0.60

Hyper

-0.59

oll

-0.59

AMI

-0.59

POSITIVE LOGITS

 ranks

1.09

 rank

0.92

 veter

0.81

 insign

0.79

ynski

0.78

cliffe

0.77

chwitz

0.76

 ranked

0.75

zik

0.75

Rank

0.74

Activations Density 0.008%

mentions of hierarchical structures or promotions within groups

the word "ranks" and its various contexts

No Comments

No Known Activations