INDEX

Explanations

phrases related to attempts or efforts

oai_token-act-pair · gpt-3.5-turbo

the word "attempt" and its various contexts in the document

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

houses

-0.79

plates

-0.71

points

-0.69

loads

-0.68

 Survivors

-0.67

faces

-0.67

 Faces

-0.66

 Sands

-0.66

 faces

-0.65

 Alps

-0.65

POSITIVE LOGITS

 emulate

1.21

 reconcile

1.18

 imitate

1.16

 recreate

1.10

 revive

1.08

 defend

1.06

 interfere

1.05

 negotiate

1.05

 penetrate

1.05

 improve

1.04

Activations Density 0.099%

phrases related to attempts or efforts

the word "attempt" and its various contexts in the document

No Comments

No Known Activations