INDEX

Explanations

terms related to procedural processes or steps

oai_token-act-pair · gpt-3.5-turbo

terms related to procedures or 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

ission

-0.77

edom

-0.75

fed

-0.73

study

-0.73

generation

-0.72

channel

-0.71

itability

-0.71

projects

-0.68

duc

-0.66

lik

-0.66

POSITIVE LOGITS

 proced

0.72

 treats

0.69

 Fist

0.68

 Boll

0.67

 Sass

0.66

 Values

0.64

 Bans

0.63

 Loll

0.63

 Properties

0.60

 Valentine

0.60

Activations Density 0.000%

terms related to procedural processes or steps

terms related to procedures or processes

No Comments

No Known Activations