INDEX

Explanations

occurrences of the word "helpful" in the text

oai_token-act-pair · gpt-3.5-turbo

expressions related to helpfulness and assistance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

metal

-0.75

gow

-0.75

buck

-0.75

thur

-0.74

Vision

-0.70

Rush

-0.69

Bam

-0.69

pool

-0.68

inction

-0.67

scar

-0.67

POSITIVE LOGITS

 helpful

0.98

 undermin

0.87

 conduc

0.78

 guiActiveUn

0.77

 useful

0.76

 aide

0.75

ãĤĭ

0.75

 behavi

0.75

 Helpful

0.74

ãĥ¼ãĥĨ

0.73

Activations Density 0.007%

occurrences of the word "helpful" in the text

expressions related to helpfulness and assistance

No Comments

No Known Activations