INDEX

Explanations

solutions or proposed actions in a given context

oai_token-act-pair · gpt-3.5-turbo

references to solutions for problems or issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 livest

-0.80

akin

-0.78

prints

-0.71

 leisure

-0.67

wood

-0.66

rew

-0.66

urt

-0.66

avia

-0.66

tein

-0.65

olds

-0.65

POSITIVE LOGITS

 solution

0.93

 solutions

0.90

 Solution

0.87

Solution

0.85

 Solutions

0.84

olution

0.82

 solve

0.79

ysis

0.76

ahime

0.71

 solving

0.70

Activations Density 0.018%

solutions or proposed actions in a given context

references to solutions for problems or issues

No Comments

No Known Activations