INDEX

Explanations

the name "Cousins" at varying activation levels

oai_token-act-pair · gpt-3.5-turbo

mentions of specific individuals and organizations, particularly sports figures and statistical references

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

oku

-0.75

ock

-0.72

dog

-0.71

Beck

-0.70

ument

-0.67

ocking

-0.64

 Friedman

-0.63

req

-0.63

ishment

-0.62

nings

-0.61

POSITIVE LOGITS

annabin

0.72

ensis

0.71

 Cousins

0.70

grave

0.70

ursive

0.69

entric

0.69

 whiff

0.68

itate

0.68

rouch

0.67

illian

0.67

Activations Density 0.047%

the name "Cousins" at varying activation levels

mentions of specific individuals and organizations, particularly sports figures and statistical references

No Comments

No Known Activations