INDEX

Explanations

names followed by a title "retired" in a text

oai_token-act-pair · gpt-3.5-turbo

mentions of individuals who are retired

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

iola

-0.71

 Grimm

-0.67

ebus

-0.65

 Hispan

-0.63

 coord

-0.63

Rig

-0.62

owitz

-0.62

hran

-0.61

 similarities

-0.61

arta

-0.61

POSITIVE LOGITS

 retiring

0.88

 retire

0.77

uting

0.75

spective

0.74

rifice

0.74

dinand

0.72

unte

0.72

ret

0.72

utes

0.70

Veter

0.70

Activations Density 0.017%

names followed by a title "retired" in a text

mentions of individuals who are retired

No Comments

No Known Activations