INDEX

Explanations

terms related to categories or groups

oai_token-act-pair · gpt-3.5-turbo

terms related to categories, sections, and commands in formal or structured contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

rov

-0.64

orthy

-0.63

AFP

-0.60

MORE

-0.60

Advertisement

-0.59

paio

-0.56

vantage

-0.56

ECA

-0.54

nova

-0.54

BIP

-0.54

POSITIVE LOGITS

 titled

0.86

 below

0.85

'[

0.84

 labelled

0.84

 shown

0.82

shown

0.82

"#

0.82

"\

0.81

'/

0.80

 above

0.80

Activations Density 0.332%

terms related to categories or groups

terms related to categories, sections, and commands in formal or structured contexts

No Comments

No Known Activations