INDEX

Explanations

phrases related to issuing warnings, statements, orders, or commands

oai_token-act-pair · gpt-3.5-turbo

phrases related to issuing or placing orders and actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ombat

-0.75

classified

-0.66

omet

-0.64

efer

-0.63

 Pegasus

-0.62

ifi

-0.62

olulu

-0.60

heimer

-0.59

 Worldwide

-0.59

behind

-0.59

POSITIVE LOGITS

redients

0.95

tons

0.80

 torches

0.77

LY

0.72

ENCY

0.72

 conduc

0.71

ãĥ£

0.70

apest

0.70

allery

0.69

 clos

0.68

Activations Density 0.084%

phrases related to issuing warnings, statements, orders, or commands

phrases related to issuing or placing orders and actions

No Comments

No Known Activations