INDEX

Explanations

words that have strong connotations or implications

oai_token-act-pair · gpt-3.5-turbo

descriptors related to how things are perceived or characterized, often focusing on the way something "sounds" or "seems."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

PT

-0.70

igers

-0.69

uve

-0.66

 walked

-0.65

 tackled

-0.64

 chose

-0.63

 eliminated

-0.63

 prefers

-0.63

atum

-0.62

 determines

-0.62

POSITIVE LOGITS

 sounding

3.68

sounding

2.38

 seeming

1.66

 smelling

1.25

 tasting

1.15

 revolving

1.02

looking

0.97

 condesc

0.87

 looking

0.87

 chorus

0.85

Activations Density 0.026%

words that have strong connotations or implications

descriptors related to how things are perceived or characterized, often focusing on the way something "sounds" or "seems."

No Comments

No Known Activations