INDEX

Explanations

phrases related to television episodes

oai_token-act-pair · gpt-3.5-turbo

references to episodic content, particularly in the context of television or series

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Poly

-0.71

 pride

-0.69

Poly

-0.67

 curve

-0.65

 widow

-0.64

 hedge

-0.64

 math

-0.63

 Splash

-0.62

 cascade

-0.61

 Fathers

-0.61

POSITIVE LOGITS

isodes

4.30

isode

1.55

issions

1.21

hemer

1.12

ovie

1.11

isons

1.09

roups

1.05

istries

1.03

leases

1.00

ku

0.99

Activations Density 0.047%

phrases related to television episodes

references to episodic content, particularly in the context of television or series

No Comments

No Known Activations