INDEX

Explanations

words related to history or past events

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate historical context

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

 Sahara

-0.78

mails

-0.71

him

-0.70

plan

-0.70

Pages

-0.68

gur

-0.68

ger

-0.68

Eva

-0.67

itas

-0.66

ity

-0.65

POSITIVE LOGITS

 housed

0.95

 conduc

0.86

 dexter

0.86

 wielded

0.84

 exting

0.83

 represented

0.83

entimes

0.82

 speaking

0.81

©¶æ

0.81

 conclud

0.80

Activations Density 0.019%

words related to history or past events

phrases that indicate historical context

No Comments

No Known Activations