INDEX

Explanations

adjectives or nouns related to size or scale

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "larger" in various contexts

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

onso

-0.75

elf

-0.70

 Exile

-0.69

ï¿½ï¿½ï¿½

-0.69

eline

-0.68

ices

-0.68

rol

-0.68

 Called

-0.67

ocracy

-0.67

SR

-0.66

POSITIVE LOGITS

 sized

1.10

 than

1.05

 intestine

1.05

than

1.01

 diameter

0.94

 quantities

0.92

 scale

0.88

 stature

0.84

 Than

0.84

 eleph

0.84

Activations Density 0.018%

adjectives or nouns related to size or scale

references to the concept of "larger" in various contexts

No Comments

No Known Activations