INDEX

Explanations

prominent or high-ranking individuals in various organizations

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ense

-0.63

trigger

-0.63

OUND

-0.63

pire

-0.63

\">

-0.62

fw

-0.61

arget

-0.61

avers

-0.61

laugh

-0.60

fitting

-0.59

POSITIVE LOGITS

 Andrew

0.99

 Theo

0.98

Ian

0.97

 Tobias

0.94

 Jamie

0.93

 Sarah

0.93

 Andy

0.93

 Patrick

0.93

 Brendan

0.93

 Betsy

0.92

Activations Density 0.210%

prominent or high-ranking individuals in various organizations

No Comments

No Known Activations