INDEX

Explanations

references to the word "telegraph."

oai_token-act-pair · gpt-3.5-turbo

references to various telegraph-related topics and names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

fare

-0.75

 Authorization

-0.70

 Duchess

-0.69

 Sergio

-0.67

stadt

-0.65

 Fitzgerald

-0.65

rior

-0.62

 Augustine

-0.62

Paw

-0.62

 Penguins

-0.61

POSITIVE LOGITS

vernment

1.09

merce

0.92

ICS

0.82

orld

0.80

pton

0.77

å¥³

0.77

legraph

0.75

itte

0.74

essional

0.72

pex

0.72

Activations Density 0.047%

references to the word "telegraph."

references to various telegraph-related topics and names

No Comments

No Known Activations