INDEX

Explanations

statistical data such as numbers and URLs

oai_token-act-pair · gpt-3.5-turbo

elements or components in scientific citations and references

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

rely

-0.81

netflix

-0.78

TextColor

-0.78

rette

-0.71

ques

-0.68

inic

-0.67

 propriet

-0.66

stood

-0.65

deen

-0.65

unct

-0.64

POSITIVE LOGITS

acity

0.96

0.84

0.80

^{

0.80

0.79

%"

0.79

0.78

0.77

Activations Density 0.077%

statistical data such as numbers and URLs

elements or components in scientific citations and references

No Comments

No Known Activations