INDEX

Explanations

formatted citations or references

oai_token-act-pair · gpt-3.5-turbo

closing parentheses or the end of expressions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ciating

-0.79

itory

-0.70

ãĤ©

-0.69

 answ

-0.69

omore

-0.68

 administrator

-0.65

 referen

-0.64

igmatic

-0.63

azon

-0.61

emonium

-0.59

POSITIVE LOGITS

 Frames

0.69

 Actions

0.68

)))

0.68

 Committees

0.67

 Immun

0.66

 Protective

0.66

 Parables

0.64

RESULTS

0.64

 Modes

0.63

 Extended

0.61

Activations Density 0.139%

formatted citations or references

closing parentheses or the end of expressions

No Comments

No Known Activations