INDEX

Explanations

references to events or incidents with a negative outcome or impact

oai_token-act-pair · gpt-3.5-turbo

instances of the word "the."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

å§«

-0.78

\-

-0.75

stat

-0.69

anton

-0.67

advertising

-0.65

hops

-0.63

dale

-0.63

emed

-0.61

alus

-0.60

stellar

-0.60

POSITIVE LOGITS

 result

1.47

 culmination

1.43

 consequence

1.13

 product

1.10

 centerpiece

1.05

 envy

1.02

 impetus

1.01

 embodiment

1.00

 catalyst

1.00

 outcome

0.99

Activations Density 0.148%

references to events or incidents with a negative outcome or impact

instances of the word "the."

No Comments

No Known Activations