INDEX

Explanations

names and historical terms related to various locations and figures

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names associated with people and locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 rollout

-0.90

 Comcast

-0.85

 analytics

-0.85

 actionGroup

-0.83

 cybersecurity

-0.82

 targeting

-0.80

 transitioning

-0.80

 broadcasting

-0.80

 targeted

-0.79

 partnering

-0.78

POSITIVE LOGITS

Ã»

1.08

onian

1.07

Ã¼

1.06

Ã¦

1.03

á¸

1.01

inus

0.98

Åį

0.97

anus

0.96

Å

0.94

idae

0.94

Activations Density 0.568%

names and historical terms related to various locations and figures

proper nouns and names associated with people and locations

No Comments

No Known Activations