INDEX

Explanations

words that are often found in code or technical documentation

oai_token-act-pair · gpt-3.5-turbo

terms related to votes or voting processes

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

Baz

-0.68

 grooming

-0.67

 pretext

-0.67

Boh

-0.64

 shrew

-0.64

Mu

-0.61

 Minute

-0.61

 Magn

-0.61

 distracted

-0.60

mog

-0.60

POSITIVE LOGITS

ES

4.05

ESA

1.76

ESH

1.73

es

1.73

ES

1.65

EST

1.64

ESE

1.61

ED

1.59

EC

1.53

EW

1.53

Activations Density 0.010%

words that are often found in code or technical documentation

terms related to votes or voting processes

No Comments

No Known Activations