INDEX

Explanations

specific company names

oai_token-act-pair · gpt-3.5-turbo

references to specific brands or companies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ãģĦ

-0.82

istically

-0.79

++++++++++++++++

-0.77

WAYS

-0.72

à¤¾

-0.69

JFK

-0.69

fare

-0.68

 ACTIONS

-0.68

plane

-0.68

~~~~~~~~~~~~~~~~

-0.65

POSITIVE LOGITS

lÃ©

1.20

 Nest

1.12

orius

1.01

ea

0.93

eus

0.86

zen

0.85

hes

0.85

arus

0.85

ensibly

0.84

eas

0.83

Activations Density 0.008%

specific company names

references to specific brands or companies

No Comments

No Known Activations