INDEX

Explanations

descriptions or explanations in a structured format, possibly in a technical or formal context

oai_token-act-pair · gpt-3.5-turbo

colons and text that indicates sections or subsections

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

ulz

-0.81

rites

-0.80

undai

-0.78

uve

-0.78

ossip

-0.75

liction

-0.72

oser

-0.71

eatures

-0.70

inement

-0.70

anyl

-0.69

POSITIVE LOGITS

 Provided

0.93

 Detected

0.93

 Unknown

0.82

TBD

0.82

 Impossible

0.80

 Continued

0.79

 Join

0.78

 Contemporary

0.78

 Programming

0.78

 Originally

0.78

Activations Density 0.098%

descriptions or explanations in a structured format, possibly in a technical or formal context

colons and text that indicates sections or subsections

No Comments

No Known Activations