INDEX

Explanations

codes or commands in a technical context

oai_token-act-pair · gpt-3.5-turbo

indicators of evaluations and classifications

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

 destro

-0.67

 dash

-0.67

versions

-0.65

":[

-0.64

 arrang

-0.63

 centrif

-0.61

 regul

-0.61

obe

-0.61

 therap

-0.59

uyomi

-0.59

POSITIVE LOGITS

asma

0.76

Unknown

0.71

ACTED

0.70

tek

0.69

Height

0.69

ixel

0.68

ructose

0.64

 Comparison

0.64

 Reply

0.63

 Requires

0.62

Activations Density 0.261%

codes or commands in a technical context

indicators of evaluations and classifications

No Comments

No Known Activations