INDEX

Explanations

instances where something is utilized or employed

oai_token-act-pair · gpt-3.5-turbo

instances of the word "use" and its variations in the context of technology or systems

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

Â¯Â¯Â¯Â¯

-0.72

antle

-0.69

hood

-0.68

worthy

-0.67

occup

-0.67

 justifies

-0.66

watching

-0.66

urther

-0.65

redo

-0.65

deal

-0.65

POSITIVE LOGITS

 borrowed

0.94

 techniques

0.87

fully

0.81

 sophisticated

0.79

 algorithms

0.79

 methods

0.78

 proprietary

0.78

 elements

0.77

GPS

0.73

 recycled

0.73

Activations Density 0.145%

instances where something is utilized or employed

instances of the word "use" and its variations in the context of technology or systems

No Comments

No Known Activations