INDEX

Explanations

the name "Brad" with varying activations

oai_token-act-pair · gpt-3.5-turbo

the repeated mention of the name "Brad."

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

 referen

-0.70

 eleph

-0.69

ktop

-0.68

 subsistence

-0.67

VALUE

-0.66

 derog

-0.66

phis

-0.65

 versa

-0.64

Magikarp

-0.63

 wiret

-0.63

POSITIVE LOGITS

shaw

1.20

enton

1.13

 Pitt

1.01

ford

0.96

 Brad

0.89

iago

0.89

street

0.85

anche

0.83

nan

0.83

bury

0.82

Activations Density 0.011%

the name "Brad" with varying activations

the repeated mention of the name "Brad."

No Comments

No Known Activations