INDEX

Explanations

words or phrases with a prefix indicating a reversal or repetition of action

oai_token-act-pair · gpt-3.5-turbo

instances of the prefix "re-" indicating restoration or repetition

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

ulhu

-1.10

 Dickinson

-0.69

 Improvement

-0.68

Nah

-0.67

 Loll

-0.65

 reviewer

-0.64

 Williamson

-0.64

 Grill

-0.63

DeV

-0.63

åŃ

-0.63

POSITIVE LOGITS

issue

1.11

introdu

1.08

elected

1.08

position

1.07

talk

1.07

tool

1.05

ded

1.02

edition

1.01

imag

1.01

branded

1.00

Activations Density 0.045%

words or phrases with a prefix indicating a reversal or repetition of action

instances of the prefix "re-" indicating restoration or repetition

No Comments

No Known Activations