INDEX

Explanations

words related to codes or names that have a specific pattern

oai_token-act-pair · gpt-3.5-turbo

terms related to codes and legal terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

esville

-0.80

athi

-0.77

swer

-0.74

kus

-0.72

 Ples

-0.71

ties

-0.71

iatus

-0.71

riks

-0.70

=-=-=-=-

-0.70

itor

-0.69

POSITIVE LOGITS

 snippet

0.98

velop

0.86

 snippets

0.83

otle

0.82

breaker

0.78

breakers

0.78

breaking

0.78

 codes

0.76

 enforcement

0.70

 reuse

0.68

Activations Density 0.028%

words related to codes or names that have a specific pattern

terms related to codes and legal terminology

No Comments

No Known Activations