INDEX

Explanations

phrases related to specific nouns or names in a context of organizational hierarchy or teamwork

oai_token-act-pair · gpt-3.5-turbo

character names and associated actions in narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 encount

-0.54

 fortun

-0.52

 Citiz

-0.52

awa

-0.52

 guiActiveUnfocused

-0.49

 NETWORK

-0.49

 Proced

-0.48

theless

-0.46

UTC

-0.46

 Smartstocks

-0.45

POSITIVE LOGITS

ologue

0.55

imei

0.53

oba

0.52

ickr

0.51

assad

0.50

kefeller

0.50

merga

0.49

EStream

0.48

astern

0.48

pai

0.47

Activations Density 1.579%

phrases related to specific nouns or names in a context of organizational hierarchy or teamwork

character names and associated actions in narratives

No Comments

No Known Activations