INDEX

Explanations

phrases indicating intensity or significance

oai_token-act-pair · gpt-3.5-turbo

phrases or concepts related to concern or issues

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

İĭ

-0.90

£ı

-0.90

nown

-0.88

Ļ½

-0.84

 sugg

-0.81

qqa

-0.81

etheless

-0.77

bris

-0.77

nect

-0.75

²¾

-0.75

POSITIVE LOGITS

ãģ®

0.67

 Britons

0.66

 Sharif

0.65

::

0.62

CPC

0.62

 redirect

0.62

 Swarm

0.62

 Guardiola

0.61

 Campaign

0.60

 sigh

0.59

Activations Density 0.080%

phrases indicating intensity or significance

phrases or concepts related to concern or issues

No Comments

No Known Activations