INDEX

Explanations

words related to visibility and legal terms

oai_token-act-pair · gpt-3.5-turbo

terms related to visibility and sight

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

ãĥ¤

-0.70

 Malaysian

-0.67

ashtra

-0.65

schild

-0.61

riad

-0.60

udder

-0.59

 Midlands

-0.59

ycle

-0.59

 Ethics

-0.59

 Cause

-0.58

POSITIVE LOGITS

cer

1.00

ially

0.97

vis

0.92

iors

0.90

 challeng

0.88

 versa

0.86

ibilities

0.81

acet

0.81

uo

0.79

ets

0.79

Activations Density 0.021%

words related to visibility and legal terms

terms related to visibility and sight

No Comments

No Known Activations