INDEX

Explanations

phrases related to European Parliament (EP) actions and topics

oai_token-act-pair · gpt-3.5-turbo

references to the European Parliament (EP)

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

 à¨

-0.73

 Schwar

-0.72

 Sunder

-0.71

ãĤ´ãĥ³

-0.69

 Sands

-0.67

 Rite

-0.66

gaard

-0.66

 Dracula

-0.64

Axe

-0.63

Peb

-0.63

POSITIVE LOGITS

EP

1.07

rint

1.05

odcast

1.02

oleon

0.98

HER

0.95

iphany

0.95

hemer

0.95

ascal

0.94

IC

0.93

resent

0.92

Activations Density 0.006%

phrases related to European Parliament (EP) actions and topics

references to the European Parliament (EP)

No Comments

No Known Activations