INDEX

Explanations

phrases describing a comparison or certain types of actions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "this" and phrases that denote examples or references

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

istries

-0.90

Ni

-0.77

erate

-0.70

half

-0.68

sent

-0.67

verning

-0.65

wa

-0.65

Wr

-0.64

ãĥ´ãĤ¡

-0.63

Fit

-0.63

POSITIVE LOGITS

 spoiled

0.73

 bookmark

0.65

 improvised

0.60

 outgoing

0.60

tip

0.60

agine

0.59

 sunrise

0.58

 modifier

0.57

 guiActive

0.57

ragon

0.57

Activations Density 0.103%

phrases describing a comparison or certain types of actions

instances of the word "this" and phrases that denote examples or references

No Comments

No Known Activations